Update from 03/31/2023, 10:03 am: We first wrote that Midjourney blocked the free access. That was not true, because according to CEO David Holz, the free trials have been suspended for the moment. They are therefore only temporarily unavailable.
It’s over and done with: Midjourney is ending free access for the time being. You have to pay for variations of Darth Vader as a hippie or the cover of the punk rock album of the band you never had, portrayed by alien cats.
The reason: As the Washington Post reports, according to CEO David Holz, it is due to the extraordinary demand and abuse of the trial version. Over time, more and better deepfakes of real people have appeared. Most recently, pictures of the arrested Donald Trump or Pope Francis in a white down jacket made the rounds.
Even if the pictures were quickly exposed as bad jokes, Midjourney suddenly saw itself in the limelight as the company that made such deepfakes possible quickly and easily.
The consequence: Free access first pull the plug and ask interested parties to checkout. This is what Holz said on the Midjourney Discord server:
An attempt to allow free access nonetheless failed, as Holz describes:
How much does mid-journey cost?
If you want to use the AI for fun, you pay at least 10 dollars a month. This buys you 3.3 hours of GPU time, i.e. the time the AI needs to calculate your requests. Midjourney’s payment plans are as follows:
|Basic Plan||Standard Plan||Pro Plan|
|Monthly charges||10 Dollar||30 Dollar||60 Dollar|
|annual costs||$96 (equivalent to $8 per month)||$288 (equivalent to $24 per month)||$576 (equivalent to $48 per month)|
|GPU time||3.3 hours/month||15 hours/month||30 hours/month|
Not the first controversy
As early as 2022, Midjourney found itself under criticism after the company admitted that it was difficult to establish guidelines for AI use.
That happened: Holz justified the ban on the use of images of Chinese President Xi Jinping via Discord. According to him, the aim was to prevent drama because it was more important to gain a foothold in China than to spread satirical images.
Is it difficult to establish strict rules? According to OpenAI, the company behind ChatGPT, no. All current political events and politicians in general are banned there, as are conspiracy theories. Also, ChatGPT prohibits hatred, sexuality and violence.
The same goes for Stable Diffusion, another image AI. There, the company prevents direct copying of certain styles or the generation of NSFW content.
Misleading content is not the only problem. The copyrights are still unclear. Of course, an AI cannot appear as the author, but references that the intelligence uses to create images are protected by copyright. A final solution is not in sight here.
Of course, artificial intelligence is not a bad thing per se. Some apps with AI support help you in everyday life or are ideal for entertainment. Presidents Trump and Biden are discussing Overwatch, for example.
Midjourney is phasing out free content due to abuse. Did you actually use the image AI? If yes, for what? Do you think that was just a logical and necessary step to prevent abuse? Or does that not deter offenders in the first place? Share your thoughts with us in the comments.